Research On The Application Of Fuzzy Theory In Medical Diagnosis And Hospital Evaluation

Medical diagnosis and evaluation is a very complicated process, and the observationsdata given by the doctor is usually fuzzy. The differentiation degree of these fuzzy data is low,which makes the diagnosing and evaluating difficult. So patient always tends to the expert incertain diseases. But the data given by the doctor is subjectivity, which cannot reflect thecondition of the patient accurately. The dependence in the various data is high, and theindicators given by the doctor are complex to deal. So a medical diagnosis and evaluationmodel is created. This model is able to assess the subjectivity, deal with the fuzzy data andreduce the number of the indicator. It focuses on the practical problems, which is verypractical.This paper created a medical comprehensive evaluation model about medical system byfuzzy theory. The overall evaluation of the probability and severity of illness will be obtainedby inputting a series of indicators. The critical problems below should be solved in this model:(1) The Measures of association. This paper calculated the value of single measures byprincipal component analysis firstly, and then calculated the measures of association bySugeno measure. (2) The ascertainment of the coincidence function. This paper obtained thecoincidence function of every indicator, which was verified by the instance. (3)Theprobability of the diagnosis could be obtained by means of the fuzzy integral. The fuzzyintegral could solve the relation of the fuzzy set and the measures of association. (4) Therelation of the severity and the probability function. This paper proposed a function, whichhad very high reference value. Then the application of the model in evaluating thecompetition ability of hospital was proved highly effective. The indicators used to evaluationare always interaction and mutual checks, so the analytic hierarchy process cannot fullyreflect the internal relationship of these indicators, and the evaluation results can not reflectthe true management level and comprehensive competition of the hospital.Main work and conclusions of this paper:1. Create a diagnosis model1）Obtain the indicators which are more important using principal componentanalysis method2）calculated the measures of association by Sugeno measure, namely the importanceof association indexes3）Ascertain the coincidence function and perform calculations based on theexperiment data4）Obtain the average figure by fuzzy integral or Choquet integral5）Ascertain the membership function and analyze the evaluating values2. Confirm the high reference value of this model by certifying the practicability andreliability of it with concrete example 1）The prediction accuracy can reach 90% when the model is used to process the testdata of Parkinson Disease2）The competition ability of private hospital can be evaluated objectively and trulybased on the various indicators by the model...